基于复合变换的湿地信息提取的决策树方法研究——以安庆沿江湖湿地为例  

On Decision Tree Algorithm of Wetland Information Extraction Based on Complex Transformation

作　　者：孙步伟[1] 赵萍[2] 程雷[1] 张旭明[1] 

机构地区：[1]安徽师范大学国土资源与旅游学院GIS重点实验室,安徽芜湖241000 [2]合肥工业大学资源与环境工程学院,安徽合肥230009

出　　处：《安徽师范大学学报（自然科学版）》2009年第3期271-276,共6页Journal of Anhui Normal University(Natural Science)

基　　金：安徽省教育厅自然科学基金项目(KJ2007B219);安徽省教育厅教学项目(2007JYXM208);安徽师范大学博士科研资助计划;安徽师范大学GIS重点学科建设项目

摘　　要：湿地具有多种独特的环境功能和生态效益,在维持地球表面生态平衡中起着极为重要的作用,因而受到人们的极大关注.因此如何快速准确的获取湿地基础信息,成为湿地研究的关键.文章以安庆沿江湖湿地为例,通过对研究区TM影像进行HSVI、HS、LBV、K-T等多种变换和采样分析,构建了新的指数和谱间关系,并在此基础上建立决策树模型提取湿地信息.结果表明,该方法能够快速、准确、客观的提取湿地信息,为湿地研究提供精确的基础数据.Being unique in various environmental functions and ecological effects, wetlands have been playing an extremely important role in keeping the ecological balance of the earth surface, and for this reason it gets great concern. How to acquire the basic information of wetlands rapidly and accurately becomes very crucial. This paper comes up with a method of getting the information of wetlands, which takes the wetlands along-Yantze and lakes in Anqing city as an example, and various transformation techniques like HSV, IHS, LBV, K - T and sample analysis were taken on the TM images of the study area and then building new indexes and spectral relationships. The decision tree model was also established. The method referred to above turned out to be rapid, accurate and objective, and can provide accurate basic data for wetlands research.

关 键 词：HSV变换 IHS变换 LBV变换 K—T变换 湿地 决策树 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]